Molecular Evolution
In 1952, Frederick Sanger and coworkers
determined the complete amino acid sequence of
insulin.
Since that time, the amount of sequence information
has grown exponentially.
For example, Genbank contains all publicly available
DNA sequences, which amounts to more than 3.8
billion basepairs from 4.8 million sequences!
In addition, the entire genomes of over thirty
organisms have been sequenced, including two
eukaryotes (the fungus, Saccharomyces cerevisiae,
and the nematode, Caenorhabditis elegans).
The human genome is also well on its way to being
sequenced, with an expected date of completion in
the year 2003.
Molecular evolution is a new field born from this
explosion of molecular information and from our
desire to understand how and why molecular
sequences have evolved to be the way they are.
Topics in Molecular Evolution
In the next two lectures, we will discuss a few
examples of research in the field of molecular
evolution:
 Detecting selection by examining
(1) Substitution rates
(2) Variability within a population
(3) Replacement versus silent changes
 Detecting historical events by examining
(1) Number of differences between sequences
 Detecting relatedness by examining
(1) Similarity among sequences within a
population
(2) Similarity among sequences from different
species
Evidence in the DNA: Selection
1. Substitution Rates
An early theme developed within this course is that,
while it is easy to establish that evolutionary change
has happened, it is difficult to establish whether
selection has played a role in this change.
DNA can provide a record of selection.
If we compare DNA sequences from different
organisms, we can estimate the rate at which
mutations appear and fix, causing basepair
substitutions.
Substitution rate = the rate at which mutant alleles
rise to fix within a lineage
(For neutral mutations, the substitution rate within a
population equals the mutation rate, since 2N
mutations appear, each with a 1/(2N) chance of
fixation.)
Variation in the rate of substitutions among regions
of the genome is due, in part, to variation in the
form of selection.
Recall that:
 A new beneficial mutation has a chance of fixing
within a diploid population of ~ 2s.
 A new neutral mutation has a chance of fixing
within a diploid population of ~ 1/(2N).
 A new deleterious mutation has almost no
chance of fixing within a large population.
Therefore, the nucleotide substitution rate is
expected to be:
 highest when mutations are beneficial,
 intermediate ( ) when mutations are neutral,
 lowest when mutations are deleterious.
Silent (or synonymous) mutations, where the
amino acid remains unchanged, are more likely to
be neutral.
Replacement (or non-synonymous) mutations
causing an amino acid change are more likely to
experience selection, but the form and strength of
selection depends on the gene and its function.
Nonsynonymous
Synonymous
Substitution
Rate
Substitution
Rate
Gene
(x 109)
(x 109)
Histone 3
0.00
3.94
Histone 4
0.00
4.52
Actin
0.01
2.92
Myosin
0.10
2.15
Insulin
0.20
3.03
Growth
hormone
1.34
3.79
Immunoglobul
in k
2.03
5.56
(From Li, 1997. Based on sequence differences
between humans and rodents, estimated to have
diverged 80MYA.)
Histones, for example, appear to have a very low
rate of replacement substitutions.
This suggests that mutations causing basepair
changes in histone genes may be deleterious.
Why?
Histones are DNA-binding proteins around which
DNA is coiled to form chromatin. Many positions
within the protein interact with the DNA or with
other histones. In addition, histones are highly
compact and alkaline.
(From Li 1997)
Most amino acid changes in histone proteins may
have negative or even disastrous consequences.
Histone proteins have strong functional
constaints.
Conversely, the amino acid sequences of
immunoglobulins (= antibody protein) evolve at a
much higher rate.
In particular, the active sites (the complementaritydetermining regions) of many immunoglobulins
actually have higher rates of replacement changes
than silent ones!
(From Ludwig Institute.)
It is thought that selection favors mutations in these
regions, thereby increasing the diversity among
antibodies produced by the body and improving the
immune response.
2. Levels of variability within a population
Another method to detect selective events is to
examine the level of variability currently present
within a population.
If a beneficial mutation appears and sweeps through
a population, what will happen to the level of
polymorphism present at neighboring DNA sites?
For example, Berry et al (1991) sequenced 1.1
kilobases of the cubitus interruptus Dominant (ciD)
locus in Drosophila.
ciD is located on a tiny fourth chromosome, which
undergoes no recombination.
They found NO VARIATION among ten D.
melanogaster sequences and only one basepair
difference among nine D. simulans sequences, even
though there were 54 differences between the
species.
By contrast, other genes from the same individuals
showed normal levels of polymorphism.
Berry et al (1991) argued that recent selective
sweeps in both species may have eliminated most of
the polymorphism on the fourth chromosome.
If there is overdominance at a site, what will happen
to the level of polymorphism present at neighboring
DNA sites?
Kreitman and Hudson (1991) sequenced a 4750
basepair region near the alcohol dehydrogenase
(ADH) gene from 11 individuals of D. melanogaster
and found higher than expected levels of
polymorphism:
(From Futuyma 1998)
There is only one amino acid polymorphism
(AdhF/AdhS) within this region, which occurs at site
1490 (see arrow).
Kreitman and Hudson (1991) hypothesize that there
is selection maintaining a polymophism at or near
this site.
ADH is an enzyme that breaks down ethanol.
Flies carrying the AdhF allele survive better when
their food is spiked with ethanol than do flies
carrying the AdhS allele (Cavener and Clegg 1981).
Nevertheless, the factors maintaining the AdhF/AdhS
polymorphism remain unknown.
3. Replacement versus silent changes
McDonald and Kreitman (1991) compared
substitutions between species and polymorphisms
within a species to construct a test to detect
selection.
Imagine that five sequences are obtained from each
of two species and that the tree relating these
sequences is:
Any mutation that happens on a red branch will
appear as a polymorphism within species 1.
Any mutation that happens on a blue branch will
appear as a polymorphism within species 2.
Any mutation that happens on a green branch will
appear as a fixed difference between species 1 and
2.
If mutations occur randomly over time and if the
chance that a mutation does or does not cause an
amino acid change remains constant, then the ratio
of replacement to silent changes should be the same
along any of these branches.
If mutations are neutral, any of these mutations has
an equal chance of being preserved.
McDonald-Kreitman test
H0 The ratio of replacement to silent changes among
polymorphic sites (within a species) should equal the
ratio among fixed differences (between species) in
the absence of selection.
If new mutations are advantageous, they will fix
rapidly and cause more fixed differences between
species.
If new mutations are deleterious, they will rarely fix,
but they will temporarily create polymorphisms.
These effects of selection should be stronger on
mutations that change the amino acid sequence
(replacement) than ones that don't (silent).
An excess of amino acid differences between
species should be seen when replacement mutations
have been beneficial and fixed by selection.
A lack of amino acid differences between
species should be seen when replacement mutations
have been deleterious and eliminated by selection.
Fixed differences
Polymorphic
sites
Replaceme
nt
7
2
Silent
17
42
Replaceme
nt
21
2
Silent
26
36
ADH gene
G6PD gene
(From Li, 1997. ADH study by McDonald and
Kreitman, 1991: 12 D. melanogaster, 6 D. simulans,
and 24 D. yakuba sequences. G6PD study by Eanes
et al, 1993: 32 D. melanogaster and 12 D.
simulans.)
For both genes, the ratio of replacement to silent
substitutions is significantly lower among
polymorphic sites within species (2 : 42 for ADH and
2 : 36 for G6PD) than among fixed differences
between species (7 : 17 for ADH and 21 : 26 for
G6PD).
The null hypothesis that selection is absent is
rejected in both cases.
The excess of replacement differences between
species suggests that mutations have been positively
favored.
Indeed, G6PD is an important enzyme in the
metabolism of pentose sugars, and it has been
argued that amino acid changes may have been
selectively favored in changing environments.
Evidence in the DNA: History
DNA sequences record other historical events
besides selection.
Sequence comparisons have been used to trace past
migration events and also past changes in
population size.
For example, Takahata et al (1995) estimated the
population size of early humans, using coalescence
theory.
Recall that all the alleles currently present within a
population are descended from a common ancestor
that lived, on average, 4N generations ago.
Similarly, any two alleles chosen at random from the
current population will share a common ancestor 2N
generations ago, on average.
These predictions also hold for two sequences of
DNA.
Therefore, if mutations occur at a rate per
generation per basepair on each of the two
branches leading to the two current sequences, the
proportion of sites that differ between the two
sequences is expected to be 4N .
Takahata et al (1995) used a more sophisticated
version of this idea to estimate N from human gene
sequences.
They studied 49 different loci from human
populations.
The total number of differences between two
randomly chosen sequences varied from zero at 37
loci to five at one locus.
Using an estimated mutation rate of 2 10-8 per
basepair, Takahata et al estimated an effective
population size for humans of 10,000.
[NOTE: There may have been more individuals alive.
10,000 represents the "effective" population size -the size of an ideal population of constant size that
would have led to the observed amount of sequence
divergence.]
Even though the current population of humans is
nearly 6 billion, the molecular sequence divergence
among humans reflects a much smaller historical
population size (~ 400,000 to 50,000 years ago).
Humans are genetically very similar, due in part
to a recent population explosion from a relatively
small number of individuals within the last few
hundreds of thousands of years.
Evidence in the DNA: Relatedness
Beyond looking for clues to our past, molecular data
can be used to tell us about the diversity within and
among species currently alive.
1. Human genetic diversity
"Accustomed as we are to noticing variations in
skin color or facial structure, we tend to assume
that the differences between Europeans,
Africans, Asians, and so on must by large...This
simply is not so: the remainder of our genetic
makeup hardly differs at all."
-- Cavalli-Sforza and Cavalli-Sforza (1995) p. 124
In a major study of human polymorphisms, CavalliSforza and collaborators studied the allele
frequencies of different populations at 110 genes.
In all cases, the differences between "races" were
quantitative not qualitative.
That is, there was not a single gene for which two
races were totally different.
Instead, slight differences in allele frequencies were
observed at most loci, e.g.:
Allele
Frequency
European African Asian
GC-1
72%
88% 76%
HP-1
38%
57% 23%
In a similar study, Nei and Roychoudhury (1982)
found that 85% of the genetic variation in the
human species exists within populations and that
only 8% is among the major "races".
"If everyone on earth became extinct except for
the Kikiyu of East Africa, about 85% of all human
variation would still be present in the reconstituted
species." -- Lewontin et al. (1984)
Similarly, Brown (1980) studied the mitochondrial
DNA from 21 humans of diverse origin.
868 nucleotide sites were examined and only a few
differences were observed between any pair of
individuals.
Overall, the sequences differed from a postulated
ancestral mtDNA sequence at only 0.18% of the
sites.
Using a substitution rate estimate of 10-8 per
basepair per year, Brown concluded that humans
passed through a severe population bottleneck
~180,000 years ago.
2. Relatedness in social insect colonies
Social insects are some of the best examples of
cooperation in the natural world.
Several queens in the wasp species Polistes
annularis found colonies together. So the queens
and workers in a colony vary in how closely related
they are. This can affect how much workers
cooperate in maintaining the colony.
Genetic data can reveal levels of relatedness and
provide key information for predicting cooperation
and conflict in these societies.
These data reveal that reproduction is not shared
equitably -- queens compete intensely for
reproductive dominance in the colony.
Sometimes a single queen produces almost all the
offspring, and the identify of that queen can change
over the lifetime of the colony.
From Peters et al. (1995) Maternity assignment and
queen replacement in a social wasp. Proceedings
Royal Society B 260: 7-12.
3. Mammalian genetic diversity
How much do we differ genetically from other
mammals?
Species
comparison
% sequence
difference
Human-Chimp
1.45
Human-Gorilla
1.51
Human-Orangutan
2.98
Human-Rhesus
Monkey
7.51
(From Li 1997. Based on 5.3 kb of non-coding DNA.)
Surprisingly little!
Tables, such as the above, provide information
about the relatedness of different species.
This data can be used to reconstruct the
phylogenetic relationships among the species
involved.
In the next few lectures, we'll discuss how
phylogenetic reconstruction is accomplished and
look at some interesting phylogenies.
SOURCES:
 Li (1997) Molecular Evolution. Sinauer
Associates, MA.
 Futuyma (1998) Evolutionary Biology. Sinauer
Associates, MA.
 Cavalli-Sforza and Cavalli-Sforza (1995) The
Great Human Diasporas. Addison-Wesley, NY.